Add editable LLM penalties and preserve sampling through backup restore

test_chat_test.py

@@ -33,6 +33,11 @@ class ChatTestsTests(unittest.TestCase):
  def test_answer_stream_and_no_prompt_in_status(self):
   self.manager.start(self.request());job=self.finish();self.assertEqual(job['state'],'complete');self.assertEqual(job['answer'],'OK');self.assertNotIn('synthetic test',json.dumps(job));self.assertNotIn('messages',job)
   self.assertTrue(self.worker.loaded);self.manager.unload();self.assertFalse(self.worker.loaded)
+ def test_profile_penalties_reach_test_worker(self):
+  self.profile['parameters'].update(repeat_penalty=1.15,presence_penalty=.5,frequency_penalty=.2)
+  self.manager.start(self.request());self.assertEqual(self.finish()['state'],'complete')
+  body=json.loads(self.worker.conn.request.call_args.args[2])
+  self.assertEqual([body[k] for k in ('repeat_penalty','presence_penalty','frequency_penalty')],[1.15,.5,.2])
  def test_memory_rejection_reaches_user_and_releases_lease(self):
   self.worker.error='GPU-Split passt nicht in den Speicher.';self.manager.start(self.request());job=self.finish();self.assertEqual(job['state'],'failed');self.assertIn('Speicher',job['error']);self.assertEqual(self.manager.scheduler.active,0)
  def test_queue_cancel_does_not_stop_other_request(self):